THE year is 1952. IBM has lifted the curtain on the 701, a world-first scientific computer, revolutionising the interface between human and machine forever.
Using magnetic strips for storage and weighing more than 20,000 pounds – just shy of ‘The Beast’ Cadillac and all its bullet-proof armouring used by the American President – the IBM 701 claimed to be the first computer capable of displaying the potential of artificial intelligence, competing with world-renowned checkers and chess players at the time to demonstrate the data-driven ingenuity hidden within this mystifying mainframe. It even managed the first English-to-Russian computer translation, eventually going on to add features such as calculators and weather forecasting – features which we now carry around in the palm of our hands.
More than 70 years later, and a timely reminder of AI’s exponential advancements arrives in the form of the fastest-growing app in history. It’s not Instagram, nor is it TikTok. Rather, it’s ChatGPT, a generative language model by OpenAI that opened to public testing in November 2022 and has since amassed a record 100 million monthly active users in just two months - a full two years faster than it took Instagram to cross the same threshold.
GPT, or Generative Pre-Trained Transformer to use its full title, is particularly adept at content creation. Perhaps alarmingly so. Its technology uses deep machine learning to create human-like text in response to simple prompts. Something akin to Siri, the AI assistant familiar to most mobile users, only strictly text-based and capable of producing more nuanced responses. The output is conversational, the potential vast.
Feed ChatGPT a few articles and it will create an essay. Submit a theme and it’ll pen a poem. Or even a joke. Its use cases range from school homework to legal research, and with its rapidly-expanding audience, ChatGPT will only become more adept – more intelligent – as new data is inputted into its feedback loop over time.
Will there come a day when website articles and blog posts feature a disclaimer to declare that no text therein was generated by ChatGPT? How about Claire Aiken, 100 per cent certified human!
But for all its capabilities, the technology is not without its faults. Already AI-assisted writing via ChatGPT has thrown up glaring concerns of plagiarism, data privacy and fabricated facts, while the application itself can only fetch data prior to the year 2021, meaning it currently has no knowledge of the past 14 months.
That will come with time, of course. The fact that Google has now unveiled its own consumer-focused AI chatbot, known as Bard, hints at the artificial intelligence arms race already in play. Rather than turn to the internet’s biggest search engine with those burning questions, people now have a new digital route to answers and explainers that’s equally accessible, albeit not as reliable.
Even Google’s AI chatbot announcement proved a costly showcase of the tech’s fallibility, where astrophysicists quickly spotted a Bard blunder relating to exoplanets and the James Webb Space Telescope. A wrong answer triggering a 9 per cent dip in stock price, equating to around £82 billion, that may well have been prevented by a human’s ability to review, contextualise and verify.
ChatGPT and its new-fangled competitors may give the illusion of sentience, but taking the human element out of communications risks removing the texture of our writing. The stories. The personality.
What it can do well – summarising any given material and drawing out the key points, for example – is free up resources for humans to focus their time on those tasks which require deeper involvement and critical thinking, which has obvious benefits to business. Working together, rather than at odds, to streamline workflows and provide a starting point on routine tasks, so long as it’s properly managed.
Generally speaking AI tends to be swiftly dismissed or even feared as a disruptor of talent, but to immediately jump to alarmist, ‘machines-taking-over-the-world’ overlooks the opportunities at its core.
Yes, there are many issues to be ironed out. And without question, the likes of Google and Microsoft will have to wrangle with data privacy concerns when betting big on AI-powered chatbots.
Chatbots which now signal a new era for human-machine interfacing. And weigh an awful lot less than 20,000 pounds.
:: Claire Aiken is managing director of public relations and public affairs company Aiken
