The post Welcome to MarTechBot appeared first on MarTech.
]]>Get answers
What do I need to know about buying a CDP?
Get creative
Write an outline of a marketing operations strategy
Explore and have fun
Write a poem about marketing operations in the style of Mary Oliver.
MarTechBot is BETA software powered by AI which will make mistakes, errors, and sometimes even invent things. Please share your feedback with us so we can improve your experience.
To help get you started, here are some best practices and sample prompts.
Best practices
Sample prompts
Are you a generative AI power user? Want to share your favorite prompts with the MarTech community? Use hashtag #MarTechBot on social media!
The post Welcome to MarTechBot appeared first on MarTech.
]]>The post ChatGPT under threat from European regulators appeared first on MarTech.
]]>There have been indications that other European regulators may swiftly follow suit. Reports suggest that France is conducting its own inquiry; Ireland has asked Italy for more details about the basis for the ban; and the German data commissioner has said that the same action could “in principle” be taken in Germany.
Why we care. Given the immense excitement created by the availability of ChatGPT and similar tools, it was perhaps too easy to overlook warnings emerging from the legal profession over the last few months that it could run afoul of European data regulations — regulations which, in many ways, have become a de facto global standard.
If the questions that arise need to work their way through the European legal system for adjudication, that could take some time, of course. But it’s clear that regulators in European nations can take swift action in the meantime.
Dig deeper: ChatGPT: A marketer’s guide
Lawful bases for processing data. One fundamental challenge for large language models like ChatGPT is that under European law, specifically the GDPR, there are only six lawful bases for processing personal data at all (data that can be used directly to identify an individual or indirectly to identify an individual in combination with other information). The bases are:
To the extent a large language model is being trained on data obtained without explicit consent, it’s by no means clear that any of these bases are applicable — unless, perhaps, one makes the bold assumption that the availability of AI solutions is in the public interest.
Data erasure. Another challenge is whether a solution by ChatGPT is competent to support the “right to be forgotten.” Under GDPR, in certain circumstances, an individual can request the erasure of their data. To be clear, ChatGPT is not scraping the web and heedlessly collecting large quantities of personal data. But it is being trained on very large sets of texts, and the question OpenAI might have to address is whether it knows what’s in those sets in terms of personally identifying information or data it might be asked to erase.
The post ChatGPT under threat from European regulators appeared first on MarTech.
]]>The post Three essentials for writing a good ChatGPT prompt appeared first on MarTech.
]]>Detail precisely what you are looking for.
Explain the context of the question: Do you need talking points? Blog ideas? If so, include that in the prompt.
Dig deeper: AI is used in marketing by two thirds of B2B orgs, Forrester finds
In my example I used the acronym CRM and spelled out account-based marketing. That’s because there are very few other uses of CRM besides customer relationship management. However, ABM has many very common uses (Anti-Ballistic Missile, Agent-Based Modeling, Activity-Based Management, etc.). I suspect ChatGPT would get the right one given the context of CRM but I wanted to be sure.
Also, one final tip: If you’re not happy with the results you got, hit the “retry” button. This tells ChatGPT to generate something different than the first results it gave you.
The post Three essentials for writing a good ChatGPT prompt appeared first on MarTech.
]]>The post Two afforable AI writing assistants in action appeared first on MarTech.
]]>WriteSonic is one of the most affordable AI writing tools designed to generate long-form pieces of up to 3,000 words. You can find helpful comments about it on the software review site G2. The kicker is that you can find pretty much the same reviews for each tool on the list, so let’s test WriteSonic on a real task.
What sets WriteSonic (WS) apart from other tools is that it prompts you to use SEO-ready templates (workflows). It’s suitable for straightforward content like “How to grow avocados at home” but is more limited for B2B pieces, where you must constantly tweak your content or ask AI to rewrite paragraphs on the fly.
That said, I’ll use a template to craft a 3,000-word article.
Step 1: Find your keywords
WriteSonic puts content optimization front and center. First, the tool asks you to specify a topic for the article and automatically puts together a list of relevant keywords. Select those you want to optimize a piece for.
Step 2: Get ideas for a title and generate an outline
The set-up process is similar to other tools, but WriteSonic limits you to choose only one tone of voice for title ideas and content generation. It cannot combine and write in different styles like Jasper or ChatGPT.
Regarding outlines, WriteSonic offers six options revolving around B2B appointment-setting services. However, I’ll go with my initial outline and see whether it can best human-written content.
Step 3: Generate a whole piece
WriteSonic did strictly follow my outline and expand on all my given points. Unfortunately, it was too repetitive and created a piece that would be easier to trash rather than edit. Read the text in the red boxes.
That’s not a usable result. But let’s give WriteSonic one more try and see what content quality we’ll get for a simple B2C piece like “How to Pick a Ripe Pineapple.“
Sadly, the tool continued to generate repetitive, slightly paraphrased paragraphs, though they certainly don’t lack sense. If I wrote a piece about pineapples, I could use this text as a rough draft and quickly extract usable ideas.
But let’s admit it, long-form content is not WriteSonic’s strong side. Jasper can do way better.
Although these comments sound like a solid “no-go” for WriteSonic, I recommend you try it for other copywriting tasks like creating social media copy or landing pages.
Copy.ai is a one-stop shop for long-form and sales copywriting. It generates full product descriptions, landing pages and emails. The system remembers your writing style and preferences.
Copy.ai has a set of features for blog post creation. It can:
Copy.ai offers two ways to write a blog post: Freestyle and Blog Post Wizard. The Freestyle mode suggests related ideas for paragraphs or sentences based on the purpose of a piece and its title. After testing this feature, I didn’t see much practical value for long-form content.
Blog Post Wizard mode follows a step-by-step process for generating blog posts. It starts by giving you an idea, follows by writing the outline and creating talking points and ends with an industry-standard article.
Craft an article with Blog Post Wizard
First, I haven’t seen a tool that would give talking points to an outline before composing a whole piece. Copy.ai has generated surprisingly solid talking points for B2B appointment setting. Check it out.
What’s more, you can generate more talking points for a heading and pick the best ones. Likewise, you can rewrite the output or add/edit a talking point.
Here’s the final draft.
This is actually good. An uncanned intro, short and on-point paragraphs without awkward or repetitive blocks. Highlighted in green is auto-completed text chosen from four Copy.ai suggestions.
I would use this copy for the rough draft without second-guessing. What’s left for a writer is to piggyback on ideas and bridge the gap with insights, real-life examples and visuals.
All tools have their target audiences, but Jasper and ChatGPT (see part one) beat any competition for blog post writing. They give you full control over composing paragraphs and rewriting, which the other tools do not.
If you are looking for a cheaper alternative, try Copy.ai. As you’ve seen in the samples, it can deliver high-quality results for one-shot articles and comes with engaging features for storytelling. And let’s not forget that Copy.ai offers unlimited words on the Pro plan.
The post Two afforable AI writing assistants in action appeared first on MarTech.
]]>The post Reply.io introduces ChatGPT-powered Jason AI appeared first on MarTech.
]]>While earlier versions of Reply.io’s technology have been available to generate email sequences, this is their first use of OpenAI’s ChatGPT.
The new Jason AI is available to existing Reply.io customers and as a 14-day trial. The pricing depends on the current Reply.io plan that users have.
Dig deeper: ChatGPT: A marketer’s guide
First emails and responses. Jason AI uses NLP algorithms and the ChatGPT language model to identify specific keywords and phrases to recognize the intention of email replies received by the marketer. The tool can then generate suggested responses.
Users can select the option to review responses before they’re sent out. If marketers and sales associates are comfortable with the AI’s performance, they can opt for the tool to automatically send out messages without review.
“Right now we employ the concept of ‘human in the loop’ where all conversations are confirmed and proofread by humans to ensure quality and security of replies,” said Anatol Kisil, Reply.io’s VP of product. “This will greatly boost teams and individuals productivity, allowing focus on more smart work instead of repetitive conversations.”
Kisil said, “Jason AI can generate the first email and the follow-ups from a short description of the user’s offer, so users can run the sequence in several minutes. It’s also creating whole conversation sequences optimized for business needs like sales, recruiting, onboarding, events, fundraising and others.”
Beyond email. Jason AI can also suggest other channels where contacts are — like SMS, LinkedIn and WhatsApp — depending on what data is available on current contacts through their previous messages, and through a CRM integration in the Reply.io platform.
“For now Jason AI can write new emails and replies to emails — more social channels will follow in the coming months,” said Kisil. “Our platform is a provider of data solutions as well, so contacts are sourced from our prospects and organization database.”
Why we care. ChatGPT integrations are flooding the B2B tech space this year. With tools from the likes of Salesforce and HubSpot, organizations at the enterprise level and smaller that use a CRM will have the opportunity to pair their data set with ChatGPT capabilities and automate some of the work that goes into their personalized outreach.
The post Reply.io introduces ChatGPT-powered Jason AI appeared first on MarTech.
]]>The post How CMOs should respond to ChatGPT’s marketing impact appeared first on MarTech.
]]>ChatGPT has been taking over every social media feed and news headline since its debut in late 2022. The tool challenges the one trait humans always thought they would have over machines — creativity.
As marketers explore this generative AI tool without restriction, CMOs must assess its impact along three key functions: content production, ideation and market research.
Vendors have been touting AI’s ability to improve creative productivity by eliminating mundane tasks for years. Still, tools like ChatGPT have raised the bar by creating draft copy and briefs on par with interns or entry-level employees — potentially making their skills redundant. Staff should be sourced and evaluated on the ability to act as editors and supervisors for machine-generated marketing copy and proposals, such as agency creative briefs.
Relieved of first-draft writing tasks, they’ll need the editorial skills to turn generic prose into distinctive expressions of brand voice and business goals while catching anomalies. As basic creative tasks become operationalized, marketing teams can shift resources to more strategic work.
ChatGPT and its visual counterparts are powerful tools for creative brainstorming. Creative teams are rapidly developing new skills in “prompt engineering.” learning how to use language in chatbot sessions with ChatGPT to issue prompts.
This skill will accelerate and broaden the scope of creative concepts that can be generated and tested, challenging traditional branding methodologies by encouraging out-of-the-box thinking that existing guidelines may not anticipate.
Dig deeper: Does ChatGPT pose an existential threat to marketers?
ChatGPT is skilled at emulating customer personas in hypothetical interviews. While caution is warranted in mistaking its answers for actual field research, it can:
In addition, brands can use ChatGPT to gather and curate competitive information across messaging, product offerings, advertising and reviews (although its model is currently based on information available prior to 2022). But this also means competitors can do the same.
Dig deeper: Three things ChatGPT needs which only you can provide
Separate from data-related challenges and the possibility of “hallucinations” and wrong answers, CMOs must recognize numerous ethical and brand-related risks when planning for generative AI’s responsible use.
Gartner expects 70% of enterprise CMOs will identify accountability for the ethical use of AI among their top concerns by 2025. Misinformation, bias, copyright and transparency are all serious considerations in the ethical use of ChatGPT in marketing.
As one example, ChatGPT, like the internet it learns from, lacks veracity. It’s trained to predict the content of an answer without evaluating its basis. This lack of truth-based context demands that brands continue to apply human oversight.
Other brand-related cautions include:
Tools like ChatGPT are also likely to increase the volume of text produced by bots in social media with ill intent. This will escalate the need for organizations to monitor and respond to false and defamatory content, perhaps via a content authenticity function Gartner expects to be present in most marketing organizations by 2027.
Currently, ChatGPT is a complementary and partial alternative to search – not a replacement – since it focuses on generative approaches to answers rather than content discovery (this document, this sentence). Keep a close eye on how the increase in data in the training set and supervised learning continues to advance new versions of GPT.
Dig deeper: ChatGPT set to shake up search
ChatGPT’s out-of-the-box ability to produce consistent results limits the variability of its output. Attention to detail and human review of generated content are necessary to ensure that empathy, cultural awareness and perspective are provided to develop novel ideas.
As ChatGPT and other generative AI tools become more mainstream, marketing organizations should act on talent, budget and strategy considerations in the nearer term.
Review staffing plans, skill sets and job descriptions to emphasize skills and experience with reviewing, editing and fact-checking text. From an agency perspective, ensure internal and external talent are up-to-speed on public AI usage policies and seek out creative developers familiar with ChatGPT API.
Free and low-cost tools like ChatGPT may be able to replace some costly features and functions supported by legacy martech applications and agency feeds, so work with IT to identify redundant capabilities and expect agency budgets for content production to drop.
Accelerated content generation will put more pressure on distinct brand positioning as copy production explodes and more competitors rely on common models. Use newfound employee capacity to create diverse internal testing and review panels to speed up the evaluation of messaging, ads, social posts and other assets before committing to distribution.
CMOs urgently need to compile a list of active use cases impacted by ChatGPT and collaborate with peers to establish policies and practical guidelines to steer its responsible use. Start with the basics as the possibilities of the software continue to get unpacked.
For many organizations, leveraging custom models under IT governance, with an eye toward marketing’s pragmatic use cases, will yield the most transformative benefits.
ChatGPT marks a pivotal turning point in the role of generative AI in business and society. It may give way to newer products over time, but its capabilities will surely transform how people interact with technology.
The post How CMOs should respond to ChatGPT’s marketing impact appeared first on MarTech.
]]>The post 5 AI writing assistants in action appeared first on MarTech.
]]>ChatGPT’s launch in November 2022 spurred a debate among marketers. Proponents of generative AI envision a future where these highly sophisticated tools will replace human writers. Some are confident in AI’s capability to deliver high-performing copy for ads, landing pages and even white papers and ebooks.
In turn, more realistic marketers and copywriters assert the stark superiority of human writing over AI. “It cannot connect the dots like we can. It can’t be personal like us, human like us. It can’t feel the gravity of a profound moment like birth. These are uniquely human things. And so, AI cannot write like us,” wrote Eddie Shleyner, founder of VeryGoodCopy.
But there is a middle ground. We can leverage generative AI tools to level up the content creation process. It takes considerable time and effort to write high-quality blog posts. Using AI tools as writing assistants can potentially save costs and resources.
Because not all tools are created equal, we tested five popular AI writing assistants today. While some results were impressive, the outputs tended to be repetitive. This is a key theme across these generative AI tools, highlighting the need for human oversight and careful editing.
Here are the five AI writing tools we’ll cover:
ChatGPT (Chat Generative Pretrained Transformer) is a high-profile natural language processing (NLP) model that enables AI-powered, real-time conversations. It is great for generating short content, like creative ads, persuasive copy, scripts and more.
But there’s a caveat to generating blog posts: you never know what outcome you’ll get, its length and completeness. Aside from that, you will have to deal with repetitive word patterns and identical sentences.
To some extent, you can bypass the limitations with creative prompts and by “writing” your piece in several rounds with follow-up questions.
Dig deeper: ChatGPT: A marketer’s guide
Step 1: Write a title
ChatGPT works great in generating creative headlines. Here’s a sample prompt to get started:
Step 2: Craft an engaging introduction and outline
ChatGPT can help here if you prompt it well, for example:
And you’ll get something like this.
Not bad for an AI tool, but it’s still raw. Your next move is to try different prompts and specify which parts you don’t like. If you stay in one chat, it will pick up your writing style and adjust its responses accordingly.
To achieve the best results, run several rounds of prompts and take the most compelling parts from each intro to create an effective opening line. Then, repeat this process to develop an outline.
You might think it’s a lot of work to write the lead — you’re right. Initially, writing with ChatGPT can eat up your time. But after some trial and error, you’ll eventually develop better prompt templates and save time.
Step 3: Generate an article
Try writing in blocks. For starters, take one heading and ask ChatGPT to expand on it. Specify the following:
For instance, I asked it to put together best practices for visuals for florist websites, and here is what I’ve got.
All the points are valid, so you can piggyback on some of them as a writer.
Note: You can also ask ChatGPT to back up its outputs by referring to research, case studies, customer success stories, etc. The caveat: it often makes up stats and research, so it’s better to leave this part to humans.
Pro tip: Make ChatGPT provide fresher data from the web, research, case studies and more with the WebChatGPT Chrome extension. (It’s also available for Firefox.)
Here’s what you can do with Writer:
Writer also tries to mimic human writing style by understanding the context behind every sentence and generating natural-sounding suggestions accordingly. It can also pick up your brand voice and style guides to ensure consistent content.
There are two ways to start crafting an article with Writer — from scratch or with an outline. Since content writers are usually given an outline to work with, let’s see how Writer can fill in the gaps.
Choose the Blog posts template and fill in the working title, summary, and headings. Click Generate content. Wait for a while and enjoy the generated article.
Unfortunately, I ended up with an article full of fluff, general statements and an overblown intro. I might salvage a few paragraphs.
Here’s the intro:
And here’s some of the body text:
The first two paragraphs were good, but the rest didn’t bring real value to the reader. In comparison, here’s a human-written version.
Easy summary
Extract the juice from your article in bullet points for a TL;DR section or social media in one click. This feature saves heaps of time in summarizing text.
Event takeaways
The Event takeaways feature is impressive. It transcribes and summarizes podcasts, webinars, and other video and audio content types in minutes.
Keep writing
Enter a sentence and prompt Writer to help you get started. I like to use this feature for generating examples or completing simple sentences. For instance, I can start with “Email marketing is …” and make Writer craft the whole paragraph.
See how it works.
Since its launch in February 2021, Jasper has become a vital tool for many writers and content creators. In less than two years, it claimed Unicorn status after a $1.5 billion valuation.
This is my preferred tool among all AI writing assistants alongside ChatGPT.
There are several ways to create an article. You can opt for template-based writing if you’re a beginner or use its advanced features to get creative and develop unique content ideas.
Tip 1: Generate a full blog post with one click
Template-based writing with the One-Shot Blog Post feature enables you to generate a blog post from an outline. Just feed Jasper the main topic (a brief), select the tone of voice and target audience. Watch Jasper craft a complete article in minutes.
I used the same prompt from the Writer example: “Write an article about B2B appointment setting.”
The following sample is what you can expect for B2B content. It serves as a good outline with talking points that still need a human writer to fill in the gaps and expand on ideas. Jasper can also help with the latter in Commands mode (more below).
Tip 2: Assemble a blog post from intro to conclusion with advanced settings
Once you have keywords and the blog post’s objective, you can turn it into an engaging B2B article with a title and an outline. To do this, head to Documents in Boss Mode (find it in Templates).
Now, choose Blog Post or Blog Post Starter from Workflows, and start guided writing with an AI assistant.
First, generate 3-5 or more intros in one click and regenerate the content if you don’t like it. Below is a generated version from the first attempt, and it’s reasonable to keep it as a skeleton for further edits.
Pro tip: Set your company’s tone of voice and audience to write as closely to your brand style as possible. You can specify several tones, like conversational, informative or persuasive.
Next, create an outline in one click. Not sure whether you should keep it? Copy and save it before commanding Jasper to try again.
Here’s the first version of the outline (see below).
You can start filling in the gaps with the tool. At this point, Jasper has already remembered the content structure and is ready to generate paragraphs for each subheading.
Switch to Focus Mode and co-write a piece using commands and auto-completion. No other AI writing tools offer these features of similar quality. Co-writing with Jasper can take the words out of your mouth.
Pro tip: Use commands to enrich your piece with relevant examples in seconds, e.g., “How to calculate ROAS” or “Create an example of a curiosity-driven outreach email to sell [your product].”
In part 2 of this two-part series, we’ll look at AI writing tools WriteSonic and Copy.ai.
The post 5 AI writing assistants in action appeared first on MarTech.
]]>The post Grammarly introduces generative AI tool GrammarlyGO appeared first on MarTech.
]]>“Generative AI represents an inflection point in innovation that Grammarly can incorporate to deliver even more value for our customers,” said Rahul Roy-Chowdhury, Grammarly’s global head of product, in a company release.
Availability. GrammarlyGO will be made available across Grammarly’s product offerings, which boast 30 million daily users globally.
Users of Grammarly Free will be able to try the new product in select markets. GrammarlyGO will also be available for Grammarly Premium, Grammarly Business, Grammarly for Education and Grammarly for Developers.
Dig deeper: HubSpot debuts ChatSpot generative AI tool
Content capabilities. GrammarlyGO allows users to compose a “high-quality” draft with a single prompt. For emails, it generates short responses like “I’m interested” with one click.
It will also use contextual clues in texts to power revisions of drafted texts. It aims to improve the “tone” of the message to make it sound more professional or more friendly and personal. Additionally, it claims to be able to adjust the length of the text with a single click.
GrammarlyGO will also provide one-click prompts to jumpstart or improve writing.
Why we care. OpenAI’s ChatGPT burst on the scene late last year, and has grown to upwards of 100 million users since then. This week, Salesforce unveiled ChatGPT integrations for sales and marketing clouds, while HubSpot kicked off an alpha version of an enhanced CRM chat assistant. This suggests that purely from a brand standpoint, more users will become acquainted with ChatGPT in the months to come, and will either adopt and trust the technology or move onto something else.
Grammarly, on the other hand, has users who’ve been working with the technology for years. These regular users already have Grammarly generating revision suggestions for them, so there’s less of a leap of faith required to try out GrammarlyGO, when it becomes available later this year.
The post Grammarly introduces generative AI tool GrammarlyGO appeared first on MarTech.
]]>The post HubSpot debuts ChatSpot generative AI tool appeared first on MarTech.
]]>This week, Salesforce also unveiled Einstein GPT, integrating ChatGPT functionality with their CRM, sales clouds and Slack.
ChatUX. The ChatSpot app aims at providing an easier way for users to work within HubSpot’s CRM using general language prompts. This technology, which Shah calls “ChatUX,” can help users interact with the CRM and perform various marketing and sales functions in fewer steps.
CRM. Shah shared several use cases for ChatSpot that applied to CRM and other use cases.
For example, if a user wants to add a contact to the CRM, all the user has to do is type this contact’s information into ChatSpot, instead of weeding through menus in the CRM.
The ChatSpot app can also provide answers to help with sales research, fielding questions about contacts that are located in specific markets, or organizations that are a specific size or are part of a specific industry.
Content. The ChatSpot app can generate emails and other messages that pull in contextual information about the company and contact the salesperson is addressing.
With a single click, the user can copy these messages into a clipboard to use in their preferred channel.
Reporting with ChatGPT data sets. While ChatSpot provides an easier natural language interface for accessing data within the HubSpot CRM, it also pulls in additional data sets through its integration with ChatGPT.
In generating a report on global markets, for instance, the user can add columns of data from ChatGPT about populations of countries and other data that isn’t in the CRM.
Why we care. This is an early look at some generative AI and chat functions that have the potential to cut down on a marketer’s repetitive tasks. As Shah emphasizes in the video below, the content that ChatSpot generates for emails is a “first draft” intended for a human to build off of and proofread.
The post HubSpot debuts ChatSpot generative AI tool appeared first on MarTech.
]]>The post ChatGPT: A marketer’s guide appeared first on MarTech.
]]>The tool quickly attracted techies and non-techies because one prompt can generate a response usable by editors, PR teams, developers or executives to create white papers, software programs, client presentations, press releases and more.
As marketing professionals, it’s critical to learn what it can and cannot do today, how best to use it for our campaigns — and what’s in store for us.
Launched by OpenAI in November, ChatGPT is built on top of OpenAI’s GPT-3 family of large language models (LLM) and enables interaction with a model via a conversational user interface.
ChatGPT computes the next most probable set of letters or words when given an initial starting phrase or “prompt.” It was trained on 300 billion words taken from books, online texts, Wikipedia articles and code libraries — reportedly using a snapshot of the internet as of 2021.
“As a language model, ChatGPT is best used for tasks that involve natural language processing, such as text generation, text completion and answering questions. Specifically, ChatGPT can be used for a variety of tasks including:
1. Chatbots: ChatGPT can be used to power chatbots and virtual assistants that can engage in natural language conversations with users.
2. Content creation: ChatGPT can generate human-like text on a variety of topics, which can be useful for content creation, such as writing articles or generating product descriptions.
3. Question-answering: ChatGPT can provide answers to questions posed by users, such as providing information about a particular topic or field.
4. Translation: ChatGPT can be used to translate text from one language to another.
5. Text summarization: ChatGPT can be used to summarize longer pieces of text into shorter, more easily digestible summaries.
6. Personalization: ChatGPT can be trained on specific datasets to provide personalized responses or recommendations based on user behavior or preferences.Overall, ChatGPT is a versatile tool that can be used in a wide range of applications that involve natural language processing.”
This section was generated by ChatGPT with the prompt: What is ChatGPT best used for?
Today, the outputs of ChatGPT are impressive in documentation areas. It can assist — but not replace — humans in writing reports, outlines, creating press releases, books or developing surveys. It helps writers and editors get started in developing pieces of content. When you have writer’s block and are stuck, ChatGPT can come to the rescue.
Dig deeper: When I asked ChatGPT to write an article about ChatGPT
As the log-in page of ChatGPT points out, it does have limitations, like any new technology. ChatGPT may occasionally generate incorrect information, including harmful instructions or biased content.
Remember, this tool is based on a snapshot of the internet in 2021, a year of pandemic mania, racism and unmitigated lying. The tool is only as good as the data that supports it. That’s why it cannot answer questions about events that happened from 2022 to date.
Below are some other areas where ChatGPT has limits:
At best, it is a complement to search and only a partial alternative. ChatGPT focuses on generative approaches to answers rather than artifact discovery (e.g., a particular document or sentence). It also does not tell you from what source it pulls the information, meaning anything it reproduces is not attributed to the original author.
ChatGPT can translate simple sentences into traditional languages. However, it can have problems with translations for domain-specific areas and languages well beyond the Mediterranean languages. A human translator should be engaged to validate the translation.
ChatGPT users should treat the information generated as a public site post and avoid publishing personally identifiable, company or client information. Users’ conversations with ChatGPT could be used for training new models and will be reviewed by trainers. You can’t delete specific prompts, so be careful what you share. While you can delete an account, this will not delete the training data.
Users should carefully evaluate the inputs and outputs of the tool for misrepresentations and biases. ChatGPT is enhanced to align with the trainers’ preferences rather than verified facts. This means that output is plausible but not reliable for many use cases. Moreover, bias might be present in the large datasets that train the model.
Here’s an example close to home. I used ChatGPT to generate a statement about my accomplishments and background. For the most part, it was very accurate, but when it had me starting my career as a software programmer, I immediately saw the tool’s limitations. Since I was in the technology business and had worked for several large tech companies, it inferred that I must have started as a software engineer. I didn’t. I started in marketing.
Dig deeper: Three things ChatGPT needs which only you can provide
Knowing ChatGPT’s strengths and weaknesses can help you decide where to use the tool. Since natural language processing (NLP) is what it does best, start with the tasks that you might need as a marketer.
Usually, half of your time developing some of these pieces is spent researching. ChatGPT will be a good go-to for research assistance if you remember the limitations above.
Here are two last pieces of advice when using ChatGPT:
ChatGPT belongs to the category of growing generative AI tools, where you can also find Dall-E, which generates digital images from prompts. In February, Google announced it would enter the market with Bard, its version of ChatGPT. From what we know of the two tools, here are the key differences:
ChatGPT (OpenAI and Bing) | Bard (Google) |
• Purpose: to predict the next word • Can be used to: — Generate sentences — Summarize • No Open Source Code • Cannot be re-trained | • Purpose: to produce abstract expression • Can be used for: — Search — FAQs — Translation • Open Source Code • Can be re-trained with your data |
As this tool category expands, expect better outputs and current references. (Google will use its own data set to train Bard). The critical question is: “Will it replace marketing writers?”
No. Like all tools, it will supplement and complement the writers’ capabilities but not replace them. After all, there is a limit to what a computer can create.
Sentient computers are still several decades away. In the interim, we need to get used to using tools like ChatGPT because the limitations of technology will always be in the hands of humans.
Dig deeper: Does ChatGPT pose an existential threat to marketers?
The post ChatGPT: A marketer’s guide appeared first on MarTech.
]]>