iOrwell: Speech-Policing AI Wins MIT Prize, but Watch for Abuse
Silicon brains will keep us all safe from uncomfortable ideas. We forget that AI is programmed by human beings with biases.

As technology advances, count on those in power to abuse it to clamp down on dissent. An artificial intelligence breakthrough at the Massachusetts Institute of Technology offers reason for vigilance to protect our liberty in the years ahead.
Inclusive.ly won second place at the universityâs Entrepreneurship Competition for AI that can âdetect words and phrases that contain bias and can measure the level of bias or inclusivity in communication.â
The algorithm promises clinical precision, ensuring nobody has to speak up, argue, or risk offense; silicon brains will keep us all safe from uncomfortable ideas. We forget that AI is programmed by human beings with biases.
Not to mention that AI going rogue is the basis of hundreds of dystopian stories from âMaximum Overdrive,â where trucks become self-aware, to âThe Terminator.â
Inclusive.lyâs says its goal is âhelping people and organizations create a more inclusive environment,â good intentions that are only a few keystrokes away from being perverted into Orwellâs âNewspeak.â
In his dystopian novel â1984,â people, not computers, police language and ensure âgoodthinkâ falls in line with the stateâs ideology, rendering dissent impossible by eliminating the language necessary for criticism.
The MBA candidates who co-founded Inclusive.ly donât seem to have given any thought to how their technology could be abused, which is a common blind spot of good, smart people, such as Alfred Nobel.
Nobel named his new invention, dynamite, after âdynamis,â the Greek word for âpower.â A pacifist, he hoped that power would be used to end war, and instead saw it employed to blow people to pieces.
One of Inclusive.lyâs co-founders said, âWeâre here to create a world where everyone feels invited to the conversation.â Sounds like a noble â or Nobel â goal, but arenât college campuses supposed to be that world?
Iâm old enough to remember when one attended university to be exposed to new ideas, to meet people from diverse backgrounds, to ask or be asked ignorant questions, to expose the bigot and uplift the marginalized, all without the aid of AI.
Requiring everyone to operate within strict intellectual parameters is the definition of discrimination and exclusion. As General George S. Patton said, âIf everyone is thinking alike, then somebody isnât thinking.â
A free society relies on people with different opinions speaking their minds, but technology can put a stop to that. We know because even without AI, college campuses are hostile to conservative thought, pro-Israel speakers, even comedians.
This is what Newspeak categorized as âwrong-think,â and far from making everyone feel included, it results in a fear to even form contrary thoughts, much less express them, on pain of retribution.
Inclusive.ly isnât yet that sort of tool. It only makes âsuggestions for improvement,â but think of Communist Chinaâs social media scores. Such a regime wonât apply this AI in a benevolent way to ensure fairness.
The regime already has a sophisticated network to police speech and its âGreat Firewall of China,â a cute Western name for the insidious regimeâs control of information. Forget ever referring to Taiwan or the Tiananmen Square massacre if the leaders apply this innovation.
The MIT students also say Inclusive.lyâs AI âcan detect discrimination, microaggression, and condescension,â terms with so much gray area that the algorithms will have to reflect the bias of the programmers.
The neologism âmicroaggressionâ is an example of wrong-think so ephemeral it can cover anything. Several moments from American presidential debates come to mind. Vice President Gore sighing in 2000. President Biden telling President Trump to shut up.
Would Inclusive.ly flag President Reaganâs 1984 quip about his opponent Walter Mondaleâs âyouth and inexperienceâ as discrimination or age bias? How about referring to Istanbul, not Constantinople; Jerusalem as Israelâs capital; or Imperial Japanâs âcomfort womenâ?
Is the preceding paragraph too condescending for the algorithm? If so, it would suggest a rewrite â a digital version of the political officers that rode herd in the Soviet Union.
The achievement of Inclusive.ly is impressive and the students have a right to be proud. Yet as a free society, we should be wary of technology that promises to think for us, because to quote the fictional spy Sterling Archer, âThatâs exactly how âMaximum Overdriveâ starts.â