Mitigating Risks or Being Honest…
I had a healthy debate about this topic with my dear sister the other day. She asked, “What’s the difference between having a speech or ghostwriter versus using Generative Artificial Intelligence (Gen AI)?” Her question got my neurons spinning. Should (can?) there be a widespread doctrine as to how people use generative AI? The answer is not only complicated, it’s complex. Not least because we’ve only just started to scratch the surface of what Gen AI will be. And then, what’s around the corner? At the end of the day, the answer to the question: “To what extent and how should I admit usage of Gen AI in any work I do?” will depend on your context, intention and ethics. Are you in the mode of avoiding embarrasment or, worse, a lawsuit? Or would you rather be honest, have integrity? In other words be “one” with your outputs. In the realm of total transparency and to model the suggested behaviour, I was intrigued to see what my favourite Gen AI tool, Perplexity, had to say about the topic. In the text below, I used Perplexity to help construct the title headings as well as some phrases within the paragraphs. I also asked it to give the final product a run through to improve it (and it gave me some good edits!). Importantly, though, I stand behind absolutely everying herein.
What should you consider?
If you’re using Gen AI to publish under your own name or brand, here are the five major criteria to assess how you might qualify your use of AI.
1/ Ethical considerations
Ethically, it is important not to mislead audiences. For example, if readers believe content is human-created when it is not, they will feel uneasy when/if they find out. This can lead to a loss of trust and damage to the creator’s reputation/brand. I recall the case of the Koko Cares scandal, where the company used AI rather than human therapists without telling its users. It’s far less acceptable for a machine to pretend to be a human than for a human to pretend to be a machine. The former is, of course, entirely germane to the Turing Test. When looking at the ethical considerations, it must start with your ethical charter — assuming your company has one. How do you consider what is the right thing to do? Secondly, it depends on (a) the nature of the content and the stakes at play, and (b) the importance of authority and authenticity.
2/ Trust and transparency
Following on from the importance of authorship and authenticity, the origination of content — if not originality — is central to ownership. If a content creator is keen to build a brand — i.e. a reputation — the core currency is trust. To establish yourself as a reliable content creator, your ability to create above-average content is central to building a good reputation. Providing transparency regarding your process and sources — even when such might diminish your original penmanship — will ensure a greater level of trust because you’re not hiding behind another person … or machine. In a prior age, knowledge was power. Now that everyone has access to endless amounts of knowledge, your ability to originate and discriminate good from bad, or true from false, will inevitably set you apart.
3/ Legal and regulatory compliance
There is doing what is right and there is doing what is legal. If you rely on the lawmakers, you’ll likely have to wait a while and, secondly, might well find regulations that are bereft of real-world realities. Depending on the sector or industry, there may already be legal and/or regulatory frameworks that require disclosure of sources, including AI-generated content. For instance, the AI Disclosure Act of 2023 in the U.S. mandates disclaimers on AI-generated content issued by the government to combat misinformation and ensure transparency. Companies in industries such as pharmaceuticals or law that use AI without being upfront are bound to get in the cross-hairs of regulatory bodies, much less lose the respect of their customers.
4/ Intellectual property and ownership
As much as the content you create is part of your marketing arm and is designed to promote your authority, you’ll stand on shaky grounds if you use Gen AI. And any such use becomes ethically questionable grounds if you fail to acknowledge its use. If your content is part of your IP, then you’ll want to have full ownership of the result, if not the process. An example where AI-generated content can do damage to your credibility is when you move to oral explanations of your content. If you’re not the original author — having ‘borrowed’ from AI — you will struggle to own the content in public in other forums. A CEO, who has his social media and blog posts written by an intern or the company’s communications agency, will struggle to sound coherent or authoritative when quizzed in a media interview about the content. This echoes the question posed by my sister about using a speechwriter. The truth is that most top offices in politics are known for having speechwriters. There is little pretense about NOT having a speechwriter. The sine qua non is that the politician fully owns the content of the speech being delivered.
5/ Fact checking and avoiding the inacurracies if not hallucinations
Up until now, Gen AI cannot be trusted to provide entirely reliable content. Certain journalists are only too quick to highlight when AI hallucinates answers, inventing out of nowhere sources and supposed facts. That’s not to say, though, that at some time in the (near) future, Gen AI will not have fact checking and source verification as an integral part of its process (see Semji, a specialised Gen AI solution for journalists that includes automated fact-checking). At the same time, it’s worth noting that one should never entirely trust anyone’s writing. Even the most rigorous organisations with gold standard codes of ethics and integrity is subject to error. This brings up a philosophical issue I have with the general discourse about AI where we tend to wish to hold it up to higher standards we hold ourselves.
Human + AI
The point here is that, at least for the near-term, it’s highly inadvisable to publish anything that is purely a product of Gen AI. As with many use cases of AI, the best result is AI + human. And, in the case of Gen AI, I strongly recommend only publishing if you are fully prepared to own the final output. This is a question of responsibility and accountability. The more your brand is important to you, the more you strive to live up to your values — such as honesty, integrity, fidelity, creativity — you’ll want to be at least measured and at best qualified in the way you use Gen AI in your content creation. Moreover, you will want to make sure that your content is tone appropriate for your intended audience. You will need to double check sources used and instances of bias. Human-led feedback and verification will be the best way to ensure your content stays on brand, or put another way: your brand stays on track.
Best Practices for Disclaimers
If you’re a content creator and wish to integrate the use of Gen AI (and I’m a proponent for keeping on trying out what’s available), there are some useful guidelines to consider:
- Clear and Concise Language: Use straightforward language that is easily understood. Terms like “AI-generated” or “generated with an AI tool” are now commonly recognized.
- Placement: Place the disclaimer in a prominent location where it’s easily visible to the audience. In this article, I stated upfront my usage, but I think it’s viable to include it at the end. Be careful, however, not to obscure it, for example, in the small print.
- Specificity: Clearly state the type and scope of AI involvement. For example, specify if the AI was used to generate the titles, the structure or the entire content. Other forms of usage include assisting in aspects like translation and grammar.
- Ownership: Declare your level of ownership of the content. In other words, it’s best to verify all the content as if it were your own creation. At the end of the day, if the content is published under your banner or brand name, you had better believe in and stand by your own material, AI-generated and AI-assisted.
In conclusion
By following these guidelines, you can responsibly leverage the power of generative AI while maintaining trust, transparency, and ethical integrity in your content creation efforts. It’s a fast-moving space and we all need to lean into this topic, especially those with influence and an ethical backbone and a solid constitution!
