Coded Bias is an amazing and important documentary about the legalities and ethical implications of facial recognition technology. The dangers of it include invasion of privacy and racial bias.

Vegan alert:
Reference to chips being planted in animals

Coded Bias might be the most timely and relevant film I see at MIFF2020 but I found the execution hamfisted and uncinematic. Makes sweeping generalisations with little evidence to back it up and it's moments of visual invention are cringeworthy (the female HAL riff is especially awful). Charismatic interview subjects, especially Buolamwini, but they deserve better than this film. I think my big problem is it eventually becomes about more than bias in software and algorithms (which is incredibly interesting, the Samantha Bee segment on it was really thought-provoking), becoming more about surveillance. The two topics are connected but the two are vastly different so the film ends up shortchanging both.

An essential and urgent work about how new tech and algorithms are perpetuating sexism and racism, contributing greatly to income inequality, and dangerously influencing politics and the global state of surveillance.

Remember: “Saying you don’t care about privacy because you have nothing to hide is like saying you don’t care about free speech because you have nothing to say.” The same applies to being invested in how big tech uses their resources against us.

Essential watch for anyone with an interest in big data, algorithms and how our data is used.

Seeing a 14 year old school kid being stopped on the street and harassed by plain clothes officers in London after being wrongly identified as a criminal by an algorithm was particularly eye opening. No algorithm can tell you what the long term impact or implications that the incident would have on that clearly innocent young person. This clearly illustrates that these systems are inherently flawed and can cause more harm than good.

Oh shit now algorithms know I watched this film on this particular date. Fuck,

And everyone thought I was crazy for not wanting facial recognition turned on my phone.

It’s an informative and compelling (if not very visually interesting) doc.

How bout you facially recognize this dick in yo mouth Jeff bezos

Track this FBI 🖕

Asks the beginnings of good questions, lacks insight and doesn't attempt at high resolution examination. Whines rather than proposes probable if not easy solutions, such as oversight reporting of AGI Tooling Sets, Camera Selection Options, Sociological Behavior Prediction and choice making off those sets, and of course the general double mindedness of the complaints from these specific talking heads. There's a lot of Evil Monkey from Family Guy style finger pointing rather than Engineering Solutions. Which I've always despised from self proclaimed experts. How can you possibly be an expert when all you do is complain without a clear plan of attack to solve the problem.

All Light, Everywhere presents similar concerns and a much more succinct unsanctimonious way. I'd recommend watching it when it releases wide instead.

well thanks, now i’m scared 
black mirror is starting to look like a documentary
didn’t 2001 a space odyssey try to warn us about this?
i couldn’t help but roll my eyes at the part she read directly from george orwell’s nineteen eighty four like really?? 🙄 

Odio la inteligencia artificial y al día de hoy algo que me da mucho miedo es la idea de los self-driving cars o peor aun trailers y considerando que la gente detrás de esta tecnología, en su puta vida a convivido o vaya a convivir con choferes dedicados a esta profesión, aún más ansiedad me da esta idea.

fun to see people I’ve interviewed on the phone as moving pictures

ok not to be a bitch but sorry this info was NOT new and pls spare me the “we want western democratic ideals” bullshit

Anyways it wasn’t that bad I’m just ticked off by the London girl and some other people they interviewed, the original MIT girl was great

I’ve been trying to write something about this for a while, and it’s hard to put into words how important the subject of A.I and face recognition is to all of us. I would recommend everyone watch Coded Bias.

Side Note: Not to be “that guy”, but I listen to a Wisecrack’s Culture Binge podcast, and one of the hosts worked for a “big tech firm”. She claimed that in conversation, a high up executive in the company boasted their social media in built A.I knew when user was suicidal. Instead of providing help and hotline numbers, the platform would push adverts for expensive products because typically people who are going to commit suicide will make a a very expensive purchase days before. I think about that everyday and it FUCKS. ME. UP.

It’s a good intro into how and why AI is biased and how dangerous this is for future development of society, but it glosses over some things and dares to end on an upbeat we-got-this-covered vibe, even though this issue is far from resolved; the real battle is actually just starting.

Frustrating in its focus on mixing messages, featuring both groups that want to reduce the amount of identification software being used by law enforcement and others who are looking to improve how this technology identifies people of cover.

By treating these people as on the same team, the movie only ends up making the idea of algorithms the central blame, going so far to say that because humans can make illogical choices, we must abstain from following a purely algorithmic path. We never approach the idea that we can create a logic for a better humanity, one that works to benefit people rather than create more control. Everything featured in this movie is entirely relevant and true, but to say that the issue here is that we are making our machines wrong rather than questioning why these are the machines we are looking to make keeps this worldview so small.

“The past dwells within our algorithms.”

There are so many fascinating points and terrifying revelations here, but I really think this would have worked better as a podcast. Covers way too many distinct biases and their impacts way too quickly and therefore doesn’t have enough time to give us a sense of what must be done to create AI that, like the AI Justice League suggests, reflects our Western democratic values.

Important viewing. Managed not to leave me feeling paralysed and powerless, as the legislature slowly wakes up to the threat algorithms pose to freedom and social progress.

Great subject matter but kinda dull.

Sie musste ein weiße Maske aufsetzen, um von der Gesichtserkennung erfasst zu werden. Dass sogar Software, aufgrund der ihr zugeführten Daten, rassistisch sein kann, zeigt "Vorprogrammierte Diskriminierung".

Doch lange bleibt man nicht bei diesem Thema, sondern macht noch viele weitere Fässer auf. Darunter unter anderem: die Einflussnahme sozialer Medien auf Wahlen, personalisierte Werbung und vor allem der Einsatz von Gesichtserkennungssoftware durch Polizeikräfte. Das ist alles interessant (bis beunruhigend), lässt aber eine klare Linie und Vision innerhalb der Narrative vermissen.

Aufgrund der hohen Fehlerrate der Algorithmen jedoch, gibt es erste Erfolge bzw. Verbote zumindest von Gesichtserkennungssoftware. Dass aber beispielsweise in den USA Lehrer von Algorithmen bewertet oder die Rückfälligkeitswahrscheinlichkeit von Straftätern per Computer ermittelt wird, daran scheint sich noch nichts zu ändern. Ein Glück, dass hierzulande nicht mal das Internet richtig funktioniert.

Was aber, wenn diese Algorithmen eines Tage wirklich fertiggestellt und voll funktionsfähig sind? Werden Verbote aufrecht erhalten bleiben? Grüßt "1984"? Man darf gespannt sein.

Damn the FBI thinks they can be like the paparazzi and shit, the government better start treating me like a movie star then.