After driverless cars, directorless boards?
By David Arendt

After driverless cars, directorless boards?

The title may seem provocative but pertinent after the recent release for testing of ChatGPT, a new cutting-edge AI chatbox*, built by OpenAI, a non-for-profit organization in which Microsoft is the main investor. ChatGPT follows previous generative AI systems able to produce text (GPT-3), images (DALL-E2) and speech (Whisper) used in marketing, code writing, design, entertainment and interpersonal communications. **Google and Meta have developed rival systems, all of which are big consumers of data and energy, which explains why this universe is restricted to big high- tech companies. Conversational AI models enable to carry on conversations with humans while maintaining context and awareness. The use of generative AI still requires human intervention at the beginning (prompting) and the end (fact checking to weed out wrong answers) of the process. It is not too far-fetched to imagine machines impersonating a board member in person and speech and analyzing and discussing board documents, the actual board member being a mere “prompter” in the background. Eventually, automated decision-making could take over and there is no need for a board meeting or even a board.

Back to the immediate present and ChatGPT which can form arguments and write convincing and coherent replies to questions by predicting the next plausible word in a sequence of words. Here are some examples:

There are several ways that AI can potentially help you be a more effective member of the board of directors:

  1.  Decision-making: AI can provide you with data-driven insights and analysis to inform your decisions and discussions. For example, AI-powered analytics tools can help you better understand the risks and opportunities facing the organization, as well as the impact of different courses of action.
  2. Communication: AI-powered tools can facilitate communication and document sharing, making it easier for you to stay connected and collaborate with other board members even when you are not physically present.
  3. Risk management: AI can help you identify and assess potential risks more accurately and efficiently. For example, AI-powered risk management tools can analyze large amounts of data to identify patterns and trends that may indicate potential risks, such as changes in market conditions or shifts in customer behavior.
  4. Compliance: AI can help you ensure compliance with regulations and laws by automating the tracking and monitoring of compliance-related tasks. For example, AI-powered tools can monitor social media posts or other forms of communication to ensure that the organization is not making inappropriate or false statements.

Overall, AI can help you be a more effective member of the board of directors by providing you with valuable insights and analysis, facilitating communication and collaboration, and helping you manage risks and comply with regulations. However, it is important to keep in mind that AI is not a substitute for human judgment and decision-making, and you should always exercise caution when relying on AI-generated insights or recommendations.

The text in italic is what ChatGPT spits out almost instantly when prompted with the following question: How can AI potentially help me to be a more effective member of the board of directors?
Here another prompt: Can AI simulate a human person in a conference call? Answer (in a split second): It is possible for AI to simulate a human being in a conference call to some extent, depending on the specific capabilities of the AI system in question.

For example, some AI systems are able to generate synthesized speech that sounds similar to a human voice, and can be programmed to respond to specific prompts or questions in a way that is similar to how a human might respond. However, these systems are typically not able to fully replicate the complexity and nuance of human conversation, and may not be able to effectively simulate a human being in a conference call in all situations.

Additionally, some AI systems are able to generate text or images that can be used to simulate a human being's presence in a conference call. For example, an AI system might be able to generate a virtual avatar that appears to be speaking or presenting in real-time. However, these systems are also limited in their ability to fully simulate a human being's presence and may not be able to effectively replicate all aspects of human communication.

Overall, while AI can potentially assist with certain aspects of conference calls, it is not yet capable of fully simulating a human being's presence in these situations.

While the last two sentences gave me some comfort that the role of active board member is not completely passé yet I tried again: It is unlikely that AI will be able to fully replace the role of an active participant in a board of directors meeting in the near future. While AI can certainly assist with certain tasks and provide valuable insights and analysis, it is not capable of replacing the human judgment, creativity, and emotional intelligence that is often required in board meetings.

AI can, however, potentially augment the role of an active participant in a board of directors meeting by providing them with additional data and analysis to inform their decisions and discussions. For example, AI-powered analytics tools can help board members better understand the risks and opportunities facing the organization, as well as the impact of different courses of action.

In addition, AI-powered tools can facilitate communication and document sharing, making it easier for board members to stay connected and collaborate even when they are not physically present.

However, it is important to keep in mind that AI is not a substitute for human judgment and decision-making, and board members should always exercise caution when relying on AI-generated insights or recommendations.

You will note three things in the answer to my further prompt (Can AI replace me as an active participant in a board of directors meeting?):

  • the likelihood of being replaced by a generative AI model is limited to the “near future”
  • the system seems to remember and build on immediately past conversations
  • who needs Google anymore for research?

I was tempted to try ChatGPT’s analytical and problem-solving capability of a real life investment decision faced by a board in prompting underlying documentation but quickly remembered that it is open system released to the members of the public for testing.

As noted in Roose’s referenced NYT article, ChatGPT isn’t perfect:

  • its answers are based on a statistical model trained on billions of examples of text pulled from all over the internet and are sometimes incorrect or faulty
  • unlike Google it doesn’t crawl the web for information on current events because data fed into it is pre-2021.

Which prompts Roose to predict that the real revolution will come with the rumored release of GPT-4 by OpenAI sometime next year!

So for the moment real person directors won’t be replaced by AI systems and actually a clever director will start using them in the ways suggested in the above ChatGPT responses to become more efficient.

But will in the future 

  • AI systems have a “conscience” and the full range of intellectual abilities that humans have-a state known as Artificial General Intelligence***
  • computer intelligence exceed that of humans-an idea known as Singularity and introduced by the American futurologist Ray Kurzweil***
  • AI systems represent an existential threat to humanity as immortalized by Stanley Kubrick in his legendary film released in 1968 (2001: A Space Odessy) and publicly stated by Stephen Hauking, then the world’s most famous scientist, in 2014***?

How far is this future and will it ever happen as predicted or feared? Are you ready to navigate as a director in this brave new AI world? The Jury is still out! Are you ready to navigate as a director in this brave new AI world? Maybe ILA can provide some guidance in its course “AI, a Journey for Directors: from fundamentals to risk management & ****ethics”. Given what you know now, this course may sell out quicker than a ChatGPT response to a prompt! 

David Arendt, 28 December 2022

*”the best artificial intelligence chatbox ever released to the general public” dixit Kevin Roose in an article published in the New York Times of December 5, 2022 “The Brilliance and Weirdness of ChatGPT”.
** cf How Generative AI is Changing Creative Work, by Thomas H. Davenport and Nitin Mittal, Harvard Business Review, November 14, 2022
*** All these points are masterly developed by Michael Wooldridge in his excellent book” The road to Conscious Machines”, published in 2020 by Penguin Random House.
****Ethics considerations and the impact on board liability are important issues not touched on in this paper!


Legal & regulatory news