By Thalif Deen
UNITED NATIONS, May 23 2023 – When the UN displayed a female robot back in February 2019, it was a peek into the future: a fast-paced, cutting-edge digital technology where humans may one day be replaced with machines and robots.
However, a joke circulating in the UN delegate’s lounge at that time was the possibility, perhaps in a distant future, of a robot– a female robot– as the UN Secretary-General in a world body which has been dominated by nine secretaries-general, all male, over the last 78 years.
Will it take a robot to break that unholy tradition?
At a joint meeting of the UN’s Economic and Social Council (ECOSOC) and its Economic and Social Committee, the robot named Sophia had an interactive session with Deputy Secretary-General Amina J. Mohammed.
But with the incredible advances on CHATGPT chatbot– the AI search engine is now capable of producing texts, articles, pitches, follow-ups, emails, speeches and even an entire book.
If the UN goes fully tech-savvy, will AI chatbot help produce the annual report of the Secretary-General, plus reports and press releases from UN committees and UN agencies?
But the inherent dangers and flaws in AI chat bot include disinformation, distortions, lies, and hate speech—not necessarily in that order. Worse still, the search engine cannot distinguish between fact and fiction.
Testifying before US Congress on May 16, Sam Altman, chief executive of OpenAI urged legislators to regulate AI.
Ian Richards, former President of the Coordinating Committee of International Staff Unions and Associations, (CCISUA) told IPS: “ AI is good at regurgitating what it finds on the internet and which has been put there by someone, whether accurate or not. It basically reproduces existing patterns.”
“However, our work has two parts,” he pointed out.
The interesting, high-value-added part involves talking to people on the ground in remote areas, gathering stories, eliminating biases and creating data from sources that are offline or unreliable. This is something AI would find difficult to do, he added.
The less interesting, low value-added part involves creating tables and charts, running repetitive calculations and formatting documents, he noted.
“If AI can take over some of the latter and give us more time to focus on the former, staff will be both more productive and happier”, said Richards, a development economist at the Geneva-based UN Conference on Trade and Development (UNCTAD).
“But let’s not get too caught up in the hype. And any staff member who relies too much on AI to produce original content will be quickly caught out,” he declared.
Last week the New York Times quoted Gary Marcus, emeritus professor of psychology and neural science at New York University (NYU) calling for an international institution to help govern AI’s development and use.
“I am not one of those long-term riskers who think the entire planet is going to be taken over by robots. But I am worried about what bad actors can do with these things because there is no control over them,” he warned.
Perhaps a future new UN agency on AI?
Meanwhile, some of the technological innovations currently being experimented at the UN include machine-learning, e-translations (involving the UN’s six official languages where machines have been taking over from humans) and robotics.
The United Nations says it has also been using unarmed and unmanned aerial vehicles (UAVs), or drones, in peacekeeping and humanitarian operations, “helping to improve our situational awareness and to strengthen our ability to protect civilians”.
Among the technological innovations being introduced in the world body, and specifically in the UN’s E-conference services, is the use of eLUNa –Electronic Languages United Nations — “a machine translation interface specifically developed for the translation of UN documents.”
What distinguishes eLUNa from commercial CAT (Computer-Assisted Translation) tools is that it was developed entirely by the United Nations and is specifically geared towards the needs and working methods of UN language professionals, says the UN.
Asked whether UN should have a role in the growing debate on AI, UN Spokesperson Stephane Dujarric told reporters May 22: “I think this is an issue that the Secretary-General has expressed extreme worry about — the lack of regulation, the lack of safeguards, especially when it comes to autonomous weapons.”
“And I think he’s been very clear on that. It’s one of the things that keeps him up at night… we should be releasing soon our latest policy paper on the global digital compact”
Referring to AI and the social media, he said: “These are things that need to be dealt with, within what we love to refer as multi-stakeholder settings, because it is clear that in this regard, the power is not solely in the hands of governments. It is very much also in the private sector. And the UN has been and will continue to try to bring all these people to the table.”
Responding to questions whether Guterres plans to convene an international conference on AI, UN deputy spokesperson Farhan Haq said: “I don’t have a meeting to announce for now, but certainly, these are part of the concerns that the Secretary-General himself has been expressing — the idea that as artificial intelligence develops, it needs to be monitored carefully and the right regulations and standards need to be put in place to make sure that this type of technology is not open to abuse”.
Asked if there is any chance that the Secretary-General might consider convening an international conference on AI, Haq said: “That’s certainly something that can be considered. Obviously, if he believes that this would be a helpful step forward, that is what he will do. But again, I don’t have anything to announce at this point.’
Speaking on condition of anonymity, a UN staffer pointed out that the UN once tried out an AI system to generate transcripts for meetings.
But in one instance, it incorrectly cited an European Union (EU) delegate talking about “Russia’s legal invasion of Ukraine” and another delegate accusing the World Intellectual Property Organization (WIPO) of creating a conflict in Northern Ethiopia.
The moral of the story is that AI has to be closely monitored and double-checked because it can produce incorrect information and distort facts and figures.
At a White House May 4 meeting of executives from Google, Microsoft, Anthropic and OpenAI, the maker of ChatGPT, US President Joe Biden conveyed mixed feelings: “What your’re doing has enormous potential– and enormous danger”
IPS UN Bureau Report