AI cannot yet out-perform the human brain

The Big Tech super race in generative AI has just got even more interesting with Microsoft putting a new optional dialogue chat box on the Bing search engine and adding an aspect of this to the latest Microsoft Edge browser. They hope it will help greatly increase their second place 3.19%, currently dwarfed by Google’s 91.8% of market share worldwide.

Google’s immediate rebuke – a promotional video of its own AI software, Bard, became an instant embarrassment, after it gave the wrong answer to the question: “What new discoveries from the James Webb space telescope (JWST) can I tell my nine-year old about?”

 

Bard’s mistake

Bard’s answer was picked up by an American astrophysicist on Twitter, who corrected the software’s assertion that JWST had taken ‘the very first image of a planet outside our solar system.’ To rub salt in the wound, Bruce Macintosh, the Director of the University of California Observatories then clarified that in fact he had taken such images 14 years before JWST was launched. The result was that more than $100bn (£82bn) was wiped off the value of Google’s parent company, Alphabet as investors bailed.

 

Breakthrough technology

Sundar Pichai, Google’s boss, once described developments in AI as “more profound than fire or electricity.” Surely, it’s only as good as the data it learns from and in the promo video’s case, it has learned inaccurate information. AI is often viewed in terms of ‘machine intelligence’ that can perform cognitive processes such as critical thinking, prioritising and decision-making. Some say these are intellectual abilities, up till now only possible in the higher functioning cortexes of the human brain.

Already AI is actively used in medical spheres (It’s accurately sped up the interpretation of mammogram data by 30 times in the US and here in the UK an AI-led prostate cancer diagnostic tool, JivaRDX, is currently undergoing trials). On the consumer front Roomba makes use of AI in a vacuum cleaner that maps and adapts as it cleans. AI in vehicles is used to execute precise movement functions, and there are now several warehouse and factory applications that are already in use such as CANVAS Technology, owned by Amazon. But do any of these demonstrate cognition or plain ignition?

 

The limits of AI

Techopedia defines AI as: ‘An umbrella term that includes any type of software or hardware component that supports machine learning, computer vision, natural language understanding, natural language generation, natural language processing and robotics.’ The emphasis on ‘natural’ is both impressive and scary, but it shouldn’t be confused with human. AI can learn logical facts from the past that have been documented. It can map common pathways and repeatedly uses algorithms to develop a sophisticated network of patterns it can convert into ‘natural-sounding’ language it can throw back at you, in this case in a chat bot.

AI cannot reason, make value judgements, anticipate, predict, collaborate, be uniquely and innovatively creative or take account of – to any human degree – a changing context within which the information it repeatedly scrapes is set. In other words, AI lacks those really vital business soft skills we humans are just so brilliant at!

To find out more about Lacewing Media and how we can help market your business contact Kate@lacewingmedia.com

 

https://www.techopedia.com/definition/190/artificial-intelligence-ai

https://www.economist.com/technology-quarterly/2020/

http://www.wired.co.uk/article/cancer-risk-ai-mammograms

https://aimagazine.com/technology/top-10-robotics-companies-utilising-ai-technology

https://gs.statcounter.com/search-engine-market-share