Biased machines may be easier to fix than biased humans

This morning’s Observer column:

One of the things that really annoys AI researchers is how supposedly “intelligent” machines are judged by much higher standards than are humans. Take self-driving cars, they say. So far they’ve driven millions of miles with very few accidents, a tiny number of them fatal. Yet whenever an autonomous vehicle kills someone there’s a huge hoo-ha, while every year in the US nearly 40,000 people die in crashes involving conventional vehicles.

Likewise, the AI evangelists complain, everybody and his dog (this columnist included) is up in arms about algorithmic bias: the way in which automated decision-making systems embody the racial, gender and other prejudices implicit in the data sets on which they were trained. And yet society is apparently content to endure the astonishing irrationality and capriciousness of much human decision-making.

If you are a prisoner applying for parole in some jurisdictions, for example, you had better hope that the (human) judge has just eaten when your case comes up…

Read on

Tony Brooker R.I.P.

Tony Brooker, the guy who developed Autocode, arguably the world’s first machine-independent programming language, has passed away at the age of 94. He did it to make one of the early computers, the Ferranti Mark 1 in Alan Turing’s lab in Manchester University, useable by human beings. There’s a lovely recording of him talking about it on the British Library site. Also, a nice obit in the New York Times.

Quote of the Day

“The truth is that these companies won’t fundamentally change because their entire business model relies on generating more engagement, and nothing generates more engagement than lies, fear and outrage.”

  • Sacha Baron Cohen speaking about Facebook, Google et. al.

Linkblog

What comes after Spotify?

Shortly after I wrote Building vs. Streaming in popped an email from Drew Austin, who was musing about what happens when a new product/service fills a void and thereby leads to the decline of whatever filled it beforehand.

Here’s the money quote:

The increasingly-maligned model of VC-funded, loss-leading hypergrowth in the pursuit of market dominance, understood another way, is a quest to create voids that matter, voids that will hurt if we let them emerge by rejecting the product currently filling them (the fissures of a post-WeWork world are at least perceptible now). In the early ‘00s, when Blockbuster died out, it was clear that something better was replacing it (there’s a nostalgic counterargument that I’m tempted to indulge, but let’s just accept this). Today, it’s more common to watch something decline without a replacement that’s clearly better. It’s easy to understand why physical media led to file-sharing and then streaming, but what comes after Netflix and Spotify? Does anyone think it’s likely to be another improvement? I don’t, and the companies’ Facebook-like pursuit of absolute ubiquity is why. Unlike the immediately-filled Blockbuster void, I fear the Spotify void. I already got rid of all my CDs. The residue of buildings and cities determines what gets built on top of them, and if we’re conscientious, we’ll build with a more distant future in mind.

Impeachment and the democracy business

Dave Winer has come up with an nice metaphor for the impeachment process:

If you think of the United States as a company, we’ve had a strategic partnership with Russia for the last three years, kind of like the one Microsoft had with IBM. Russia is analogous to Microsoft. They’re about to roll over us in the 2020 election. Our last gasp is the impeachment.

[…]

Impeachment is like IBM shipping OS/2 and the Micro Channel Architecture. Both were designed to rid IBM of Microsoft once and for all. But it didn’t work. It was too little too late. Microsoft came out with Windows 3.0, and IBM became a global consulting company. The company that dominated the computer business left the computer business. With the US and Russia analogy substitute “computer business” with “democracy business.”

Ouch! Full disclosure: I was foolish enough to fall for IBM’s ploy. On a research budget I bought an IBM PS2 computer running OS2. It was a turkey with only one good point: a really nice keyboard!

Building vs. Streaming

Every Saturday morning for as long as I can remember, BBC Radio 3 has had a programme at 9am called “Building a Library”, in which a group of experts review recordings of classical music with a view to recommending the one(s) that the listener should contemplate adding to his or her ‘library’. The implicit model is that the music comes on a disc, which made complete sense in the pre-streaming era. The fact that the channel is still running the programme suggests that lovers of classical music still buy discs, which I guess really marks them out nowadays from lovers of pop, rap, etc., most of whom probably get their music from streaming sources. In which case a ‘library’ is now a playlist, I guess.

On the other hand…

Further to Andrew Sullivan’s essay, discussed below, here’s an alternative interpretation from Kenneth Armstrong:

There has been some speculation that the political freedom the Prime Minister will enjoy will allow him to reveal his “true self” and to pursue a policy on Brexit that is less beholden to the extreme Eurosceptic elements in his own party. The argument then runs that we could be looking at a much less “hard” Brexit than the rhetoric might have suggested. For two reasons this assumption looks doubtful.

The first is that the idea that there is a different Boris Johnson waiting to be revealed doesn’t have an obvious basis. That he is prepared to adopt different positions to suit circumstances and to seek political advantage is a better characterisation of his pollical demeanour. That could, of course, mean that the Prime Minister might seek to adjust his Brexit strategy in light of the new political realities but that is not the same as the revelation of an intrinsic and authentic political credo hitherto constrained by prevailing political conditions.

The second reason for doubting a less “hard” Brexit is to do with the problematic characterisation of different choices as “hard” or “soft”. This language suggest a continuum of possibilities with positions hardening or softening. In reality, the choice the Prime Minister will have to make will be much more binary. It’s a choice between a loose “free trade” discipline in which a future EU-UK relationship will be better than World Trade Organisation terms but significantly reduced compared to those of EU membership, and a “free movement” discipline in which the future relationship entails a pre-commitment to regulatory alignment with the EU as a condition for maintaining market access on terms similar to those currently enjoyed. There seems little reason to believe that a Johnson Government will shift from a free trade approach to a free movement approach anytime soon.

So, as Ken says, “ Getting Brexit Done turns out to be Getting Brexit Started.”

Boris: an alternative view

Andrew Sullivan, who was at Eton and Oxford with Boris Johnson, wrote a lengthy (and contrarian) essay about him for New York Magazinebefore the election. It’s worth reading in full, especially now, but here’s the bit that struck me:

He has done what no other conservative leader in the West has done: He has co-opted and thereby neutered the far right. The reactionary Brexit Party has all but collapsed since Boris took over. Anti-immigration fervor has calmed. The Tories have also moved back to the economic and social center under Johnson’s leadership. And there is a strategy to this. What Cummings and Johnson believe is that the E.U., far from being an engine for liberal progress, has, through its overreach and hubris, actually become a major cause of the rise of the far right across the Continent. By forcing many very different countries into one increasingly powerful Eurocratic rubric, the E.U. has spawned a nationalist reaction. From Germany and France to Hungary and Poland, the hardest right is gaining. Getting out of the E.U. is, Johnson and Cummings argue, a way to counter and disarm this nationalism and to transform it into a more benign patriotism. Only the Johnson Tories have grasped this, and the Johnson strategy is one every other major democracy should examine.

None of my friends is likely to agree with this. But I found it interesting. It’s the first piece I’ve read that seemed to take Johnson seriously. And — given that he is the Prime Minister of an elective dictatorship with a large Parliamentary majority — we’d better start to take him seriously. Besides, Andrew Sullivan is a perceptive and intelligent liberal.

CORRECTION. Thanks to Kevin Horgan who pointed out that Sullivan didn’t go to Eton. Wikipedia thinks Sullivan is a Oakeshottian conservative. That’s not how I read him.

What just happened?

Insightful reaction by the Economist:

Five years ago, under David Cameron, the Conservative Party was a broadly liberal outfit, preaching free markets as it embraced gay marriage and environmentalism. Mr Johnson has yanked it to the left on economics, promising public spending and state aid for struggling industries, and to the right on culture, calling for longer prison sentences and complaining that European migrants “treat the UK as though it’s basically part of their own country.” Some liberal Tories hate the Trumpification of their party (the Conservative vote went down in some wealthy southern seats). But the election showed that they were far outnumbered by blue-collar defections from Labour farther north.

This realignment may well last. The Tories’ new prospectus is calculated to take advantage of a long-term shift in voters’ behaviour which predates the Brexit referendum. Over several decades, economic attitudes have been replaced by cultural ones as the main predictor of party affiliation. Even at the last election, in 2017, working-class voters were almost as likely as professional ones to back the Tories. Mr Johnson rode a wave that was already washing over Britain. Donald Trump has shown how conservative positions on cultural matters can hold together a coalition of rich and poor voters. And Mr Johnson has an extra advantage in that his is unlikely to face strong opposition soon. Labour looks certain to be in the doldrums for a long time (see Bagehot). The Liberal Democrats had a dreadful night in which their leader, Jo Swinson, lost her seat.

What’s happened to Labour, in other words, looks awfully like what happened to the Democrats in the US in the 1970s and 1980s.