Are artificial intelligences mature enough to seduce the market? John Leroy Hennessy, president of Alphabet, co-Turing Award winner and Silicon Valley pioneer, thinks chatbots will be sufficiently relevant to mainstream needs in a year or two.
AI, a still experimental technology?
With the success of ChatGPT launched by OpenAI at the end of 2022, the tech giants have suddenly accelerated their timetable for deploying AI within their products. By working on integrating the agent into its Bing search engine, Microsoft has further motivated the projects of its competitors. Google reacted quickly by unveiling its own tool called BARD, which it plans to integrate into Google Search soon. A hasty announcement for a product still in the development stage. After a first ball from the AI and the general disappointment with new announcements, the employees of the firm have severely criticized internally the management of the file.
For John Leroy Hennessy, Google is not behind on the fundamentals of artificial intelligence. On the other hand, it highlights the shortcomings concerning the process of transforming the concept into a real finished product, ready to be adopted by users. “As a demonstration vehicle”Google’s AI is “a great piece of technology”but the Mountain View firm did not expect to have to market its program so soon, said the president of Alphabet during a summit organized by Celesta Capital, Monday February 13, 2023, and relayed by CNBC.
A call for more caution
Google’s response may have been delayed because of the ethical pressures weighing on the firm, further justifies Hennessy. “You don’t want to put in place a system that sometimes says the wrong or toxic things”, he recalls. Leader in online search, Google is torn between the need to innovate in order not to sink and the need to develop successful and coherent products. As a whole, the technology industry must today “be more careful about the impact it has on civil society”judges the former president of Stanford University.
Finally, John Hennessy predicts that the advent of generative artificial intelligence is not yet for today. It will take a few more years to find a real use for everyone. “I think these models are just getting started. You have to figure out how to integrate them into a product flow, and do it in a way that takes into account the issues of accuracy and toxicity”he analyzes.