Episode Transcript
Transcripts are displayed as originally observed. Some content, including advertisements may have changed.
Use Ctrl + F to search
0:00
Are you working way too hard for way
0:02
too little? There's never been a better time to
0:04
consider a career in IT. You could enjoy
0:06
a recession resistant career and a rewarding
0:08
field with plenty of growth opportunities and
0:11
often flexible work environments. Go
0:13
to mycomputercareer.edu and
0:15
take the free career evaluation. You
0:18
could start your new career in months, not
0:20
years. Take classes online or
0:22
on campus and financial aid is available to
0:24
qualified students, including the G.I. Bill.
0:26
Now is the time. Mycomputercareer.edu.
0:40
It's Monday, the 18th of September, 2023. I'm
0:43
Adam Roberts, the Economist's digital editor.
0:46
Welcome to Editor's Picks, where you can hear three
0:48
highlights from the weekly edition of The Economist,
0:51
read aloud.
0:53
We have two cover stories this week. One
0:55
of them examines how artificial intelligence
0:58
might supercharge progress in science.
1:01
Amidst concerns over the potential fallout
1:04
of rapidly advancing AI, scientists
1:07
find hope in how this technology
1:09
could revolutionise the way research
1:11
is done.
1:12
Our other cover story strikes a
1:14
more sombre tone as we look at the rise
1:17
of an emboldened hard right. Europe's
1:20
populist right-wing parties are growing, in
1:22
both strength and their ability
1:24
to wield political power. That
1:27
presents a real challenge to democracy
1:29
on the continent. And
1:31
finally, we turn to America, where
1:34
the new weight-loss drugs that have been taking the world
1:36
by storm look promising for treating
1:39
alcoholism.
1:41
The stories you're about to hear are just a sample
1:43
of what's on offer in the weekly edition of The
1:45
Economist. With a subscription,
1:48
you can read or listen to all
1:50
of what we do. If you're not
1:52
already a subscriber, get a month's worth
1:54
of our digital content for free by going
1:56
to economist.com slash podcastoffer.
1:59
The link is in the show notes. First
2:08
up, how AI can revolutionize
2:11
science. Debate about
2:13
artificial intelligence, or AI,
2:16
tends to focus on its potential
2:18
dangers. Algorithmic bias
2:20
and discrimination, the mass destruction
2:23
of jobs, and even, some say, the
2:25
extinction of humanity. As
2:28
some observers fret about these dystopian
2:31
scenarios, however, others are focusing
2:33
on the potential rewards. AI
2:36
could, they claim, help humanity solve
2:38
some of its biggest and thorniest problems.
2:41
And they say AI will do this
2:43
in a very specific way, by
2:46
radically accelerating the pace of scientific
2:48
discovery, especially in areas such
2:51
as medicine, climate science, and green
2:53
technology. Luminaries
2:56
in the field, such as Demis Hassabis
2:58
and Jan Lekoun, believe that AI
3:00
can turbocharge scientific progress,
3:03
and lead to a golden age of discovery.
3:07
Could they be right? Such claims
3:09
are worth examining, and may provide a useful
3:12
counterbalance to fears about large-scale
3:14
unemployment and killer robots. Many
3:17
previous technologies have, of course, been
3:19
falsely hailed as panaceas. The
3:22
electric telegraph was lauded in the
3:25
1850s as a herald of world peace, as
3:28
were aircraft in the 1900s. pundits
3:31
in the 1990s said the internet
3:33
would reduce inequality and eradicate
3:35
nationalism. But the mechanism
3:38
by which AI will supposedly
3:40
solve the world's problems has a stronger
3:42
historical basis, because there
3:44
have been several periods in history when
3:46
new approaches and new tools did
3:49
indeed help bring about bursts
3:51
of world-changing scientific discovery
3:53
and innovation. In
3:55
the 17th century, microscopes
3:58
and telescopes opened up. new
4:00
vistas of discovery and encouraged
4:02
researchers to favor their own observations
4:05
over the received wisdom of antiquity,
4:08
while the introduction of scientific journals
4:10
gave them new ways to share and publicize
4:13
their findings. The result
4:15
was rapid progress in astronomy,
4:18
physics, and other fields and new
4:20
inventions from the pendulum clock to the
4:22
steam engine, the prime mover of the
4:24
Industrial Revolution. Then
4:27
starting in the late 19th century, the establishment
4:30
of research laboratories, which brought
4:32
together ideas, people, and materials
4:35
on an industrial scale, gave
4:37
rise to further innovations such as artificial
4:40
fertilizer, pharmaceuticals, and
4:42
the transistor, the building block of the computer.
4:46
From the mid-20th century, computers
4:48
in turn enabled new forms
4:50
of science based on simulation
4:52
and modeling, from the design of weapons
4:55
and aircraft to more accurate weather
4:57
forecasting. And the computer
5:00
revolution may not be finished yet. As
5:02
we report in a special science section,
5:05
AI tools and techniques are
5:07
now being applied in almost every field
5:09
of science. Though the degree of adoption
5:12
varies widely, 7.2% of
5:15
physics and astronomy papers published in 2022
5:18
involved AI, for example, compared
5:20
with 1.4% in veterinary science. AI
5:25
is being employed in many ways. It
5:27
can identify promising candidates
5:29
for analysis, such as molecules
5:31
with particular properties in drug discovery,
5:34
or materials with the characteristics
5:36
needed in batteries or solar cells.
5:39
It can sift through piles of data, such
5:41
as those produced by particle colliders
5:44
or robot telescopes looking for patterns.
5:47
And AI can model and analyze
5:50
even more complex systems, such
5:52
as the folding of proteins and
5:54
the formation of galaxies. AI
5:57
tools have been used to identify new antibiotics.
6:00
robotics, reveal the Higgs boson,
6:02
and spot regional accents in
6:04
wolves, among other things. All
6:07
this is to be welcomed, but the journal
6:10
and the laboratory went further still.
6:12
They altered scientific practice
6:15
itself and unlocked more
6:17
powerful means of making discoveries by
6:19
allowing people and ideas to mingle
6:22
in new ways and on a larger scale.
6:25
AI2 has the potential
6:28
to set off such a transformation. Two
6:31
areas in particular look promising. The
6:33
first is literature-based discovery,
6:36
or LBD, which involves analyzing
6:39
existing scientific literature using
6:41
chat-GPT-style language
6:43
analysis to look for new hypotheses,
6:46
connections, or ideas that humans
6:48
may have missed. LBD
6:51
is showing promise in identifying new
6:53
experiments to try and even suggesting
6:55
potential research collaborators. This
6:58
could stimulate interdisciplinary
7:00
work and foster innovation at the boundaries
7:03
between fields. LBD systems
7:05
can also identify blind spots
7:08
in a given field and even predict future
7:10
discoveries and who will make them. The
7:13
second area is robot
7:15
scientists, also known as self-driving
7:18
labs. These are robotic systems
7:21
that use AI to form new
7:23
hypotheses based on analysis
7:25
of existing data and literature and
7:28
then test those hypotheses by performing
7:30
hundreds or thousands of experiments
7:33
in fields including systems biology
7:35
and material science. Unlike
7:38
human scientists, robots are less
7:40
attached to previous results less driven
7:43
by bias and crucially easy
7:45
to replicate. They could scale
7:47
up experimental research, develop
7:50
unexpected theories, and explore
7:52
avenues that human investigators might not
7:54
have considered. The idea
7:57
that AI might transform scientific
7:59
practice practice is therefore feasible. But
8:02
the main barrier is sociological.
8:05
It can happen only if human
8:07
scientists are willing and able to use
8:10
such tools. Many lack
8:12
skills and training, some worry about
8:14
being put out of a job. Fortunately,
8:17
there are hopeful signs. AI
8:19
tools are now moving from being pushed
8:22
by AI researchers to being embraced
8:24
by specialists in other fields. Apps
8:28
and funding bodies could help by
8:30
pressing for greater use of common standards
8:33
to allow AI systems to exchange
8:35
and interpret laboratory results
8:38
and other data. They could also
8:40
fund more research into the integration
8:42
of AI smarts with laboratory
8:45
robotics and into forms of
8:47
AI beyond those being pursued
8:49
in the private sector, which has bet nearly
8:51
all its chips on language-based
8:54
systems like chat GPT. These
8:57
fashionable forms of AI, such
8:59
as model-based machine learning, may
9:02
be better suited to scientific tasks such
9:04
as forming hypotheses. In 1665,
9:09
during a period of rapid scientific
9:11
progress, Robert Hooke, an
9:13
English polymath, described
9:15
the advent of new scientific instruments
9:17
such as the microscope and telescope as the
9:20
adding of artificial organs to
9:23
the natural. They let researchers
9:25
explore previously inaccessible
9:27
realm and discover things in new
9:29
ways with prodigious benefit
9:32
to all sorts of useful knowledge.
9:35
For Hooke's modern-day successes, the
9:37
adding of artificial intelligence to
9:39
the scientific toolkit is poised to do
9:42
the same in the coming years with
9:44
similarly world-changing results.
9:52
Live deeper on AI and Science on our Babbage
9:54
Podcast.
9:55
Last week, my colleague, Alok Jha, interviewed
9:58
the technologist Mustafa
9:59
Suleiman.
10:01
On Wednesday, Alok will be asking whether
10:03
AI can change the paradigm of science.
10:06
Find Babbage from The Economist every Wednesday,
10:09
wherever you listen.
10:13
Welcome to Breeze Line, where
10:16
next-level internet speeds mean next-level
10:18
productivity. Whether it's back to school,
10:20
back to work, or back to reality, don't
10:22
let slow internet slow down your game. Kick
10:25
it up a notch with a game-changing offer
10:27
of 500 megabits per second of lightning
10:30
fast speed for only $39.99 per
10:32
month. Choose Breeze Line and get
10:34
next-level internet and faster speeds
10:37
backed by a fiber-powered network. Terms
10:39
and conditions apply. Go to BreezeLine.com
10:42
to learn more.
10:47
This is Editor's Picks from The Economist. Next,
10:50
the real threat from Europe's hard right.
10:54
A spectre is haunting Europe.
10:57
The spectre of a rising hard
10:59
right. In Germany,
11:01
the overly xenophobic Alternative
11:04
for Germany, or ASD,
11:07
has surged to become the country's second
11:09
most popular party. Its
11:12
success is polarising domestic
11:14
politics, and it seems poised
11:17
to triumph in state elections in
11:19
the East next year. In
11:22
Poland, the ruling Law and Justice
11:24
Party is leading the Poles ahead of
11:26
a general election on October 15th, and
11:30
it is being drawn further to the right
11:32
by an extreme new party, Confederation.
11:36
As we explain
11:38
in this week's briefing, there could be
11:40
more grim news to come. Next
11:44
year, the hard right could gain more
11:46
sway in elections for the European
11:48
Parliament, due to be held in
11:50
June. Marine Le Pen,
11:53
the leader of National Rally, could
11:55
win the presidential election in France
11:58
in 2027. If
12:01
she did, France would become the
12:03
second big country to be run by
12:05
the hard right,
12:06
after Italy, where
12:08
Georgia Maloney and her brothers
12:10
of Italy took power last year in
12:12
a coalition with the Nativist League.
12:14
Make no mistake,
12:17
Europe is not about to be overrun by
12:20
fascists in a repeat of the 1930s. But
12:24
the new right-wing wave presents
12:27
a big challenge. Handled
12:29
badly, it could toxify
12:31
politics, disenfranchise
12:33
a large share of voters, and
12:36
prevent crucial reforms of the European
12:38
Union, or EU. Rather
12:41
than trying to exclude hard right
12:44
parties entirely from government and
12:46
public debate, the best response
12:49
is for mainstream parties to engage
12:51
with them, and on occasion
12:53
do deals with them. If
12:56
they have to take some responsibility
12:58
for actually governing, they
13:00
may grow
13:00
less radical.
13:03
Europe's hard right has enjoyed
13:05
several surges over the past quarter
13:07
of a century. In 2000,
13:11
Jörg Heider, an anti-establishment
13:14
demagogue, shocked the continent
13:16
by entering government in Austria. His
13:18
Freedom Party is now the most
13:21
popular there. A
13:23
migration crisis in 2015,
13:26
when over one million people from
13:28
poor and war-torn countries crossed
13:31
the EU's borders, led to
13:33
another wave of support for xenophobic
13:36
and Eurosceptic parties, including
13:39
Britain's Brexiteers.
13:42
The new wave that is breaking is different
13:44
in three ways.
13:46
First, the hard right has
13:48
opportunistically found new topics
13:51
to drum up fury about. Such
13:54
parties are still anti-foreigner,
13:57
but having seen Britain's experience
14:00
Some have moderated their hostility
14:03
to EU membership and
14:05
fewer want to ditch the single currency.
14:08
All are animated by new concerns,
14:11
most obviously hostility to pro-climate
14:14
policies, which they argue
14:17
are an elitist stitch-up that
14:19
will fleece ordinary people. In
14:21
Germany, the AFD has successfully
14:24
mobilised opposition to a government push
14:27
to require people to install
14:29
expensive heat pumps in their homes,
14:32
forcing the government to water down the
14:34
measures.
14:36
The second shift is the breadth
14:38
of their support. Our
14:41
calculations show that 15 of
14:43
the EU's 27 member countries
14:46
now have hard-right parties,
14:49
which have support of 20% or
14:51
more in opinion polls, including
14:54
every large country bar Spain,
14:57
where the nationalist Vox did badly
15:00
in July's elections. Almost
15:03
four-fifths of the EU's population
15:06
now live in countries where the hard-right
15:08
commands the loyalty of at
15:10
least a fifth of the public.
15:14
The final shift is that the stakes
15:16
have been raised, particularly
15:18
at a European level. The
15:21
war in Ukraine has created
15:23
a pressing need for the EU
15:25
to welcome new members in the East,
15:28
ultimately including Ukraine. In
15:31
tandem, it will need to streamline
15:34
decision-making to reduce the
15:36
veto power member states wield.
15:39
The presence of a larger bloc
15:42
of anti-immigrant nationalists
15:44
could make that crucial task far
15:46
harder. Hungary's Viktor
15:49
Orban, a guru to other populist
15:52
nationalists, has consistently
15:54
tried to block EU reform. Imagine
15:58
if he gains more allies.
16:01
How should centrist voters and parties
16:04
respond to the threat from the hard right?
16:08
The old answer was to erect a cordon
16:10
sanitaire. Mainstream
16:13
parties refused to work with the insurgents.
16:16
Mainstream media refused to
16:18
air their views. That
16:21
approach may have run out of road.
16:23
In places, it is becoming counterproductive.
16:27
In Germany, the isolation of the AFD
16:30
has reinforced its narrative
16:32
of being the only alternative to a
16:35
failed establishment. Mainstream
16:38
parties cannot pretend forever
16:41
not to hear the voice of 20% of
16:43
voters without eventually corroding
16:46
democracy.
16:48
Meanwhile, there is more evidence that
16:50
hard right parties in Europe tend
16:53
to moderate their views when they have
16:55
to take responsibility for governing.
16:59
Exhibit A is Ms Maloney,
17:02
the first hard right Prime Minister
17:04
of a Western European country since
17:06
the Second World War. Despite
17:09
liberal fears, she has not,
17:12
or at least not yet, picked
17:14
fights with upended
17:16
migration policy or
17:19
restricted abortion or gay rights.
17:22
She has remained a supporter of NATO
17:24
and Ukraine, by no means
17:27
a given on the hard right. In
17:30
the Nordics, a similar pattern has
17:32
played out. The Finns and
17:34
the Sweden Democrats, two
17:36
nationalist parties, have become
17:39
more pragmatic since either joining
17:41
or agreeing to support a governing coalition.
17:46
Any decision to include a hard right
17:48
party in local or national government
17:51
should be taken with extreme caution,
17:54
especially in places where a history
17:56
of fascism arouses acute
17:58
sensitivity. Some
18:00
rules of the road may help. One
18:03
is that to be considered, any
18:06
party must agree to renounce violence
18:08
and respect the rule of law. Just
18:11
as important is the constitutional
18:13
context. At what
18:15
level of government should they be included?
18:19
What are the checks and balances
18:21
created by the electoral system and
18:23
other institutions? It
18:26
may make sense to allow the AFD
18:29
to take part as junior members
18:31
of local government coalitions in
18:33
Germany, for example.
18:35
It would be a disaster if the hard right
18:38
were to win France's presidency
18:40
with its enormous powers.
18:44
Last, mainstream parties
18:46
must accept that they have not done
18:48
enough to satisfy a large
18:50
and angry minority of their citizens.
18:54
Trying to accelerate the green transition
18:57
by loading people up with costs
18:59
they cannot afford. Germany's
19:01
rules on boilers, for instance, or
19:04
Emmanuel Macron's ill-fated
19:06
attempt to increase taxes
19:08
on fuel is just making
19:11
greenery unpopular. Better
19:14
communication and compensation
19:16
for the worst hit are both essential.
19:20
Failing to control national borders
19:22
alienates people, whereas
19:25
a well-managed migration system
19:27
could be shown to benefit them. The
19:30
new success of the hard right in Europe
19:33
is in part a failure of the centre,
19:36
so the centre needs to raise
19:38
its gain.
19:45
And finally, the promise of using drugs
19:48
to treat alcoholism.
19:50
impact
20:00
on their lives. Over 140,000 die
20:02
from alcohol-related causes each year. Alcohol is the
20:09
fourth biggest cause of preventable
20:11
death in America. These
20:13
statistics are especially devastating
20:16
because for many people the suffering
20:18
could have been avoided. For
20:21
decades, drugs have existed
20:23
to help with alcohol addiction. Dissulfirum,
20:27
also known as Antabuse, was
20:29
approved by the Food and Drug Administration,
20:32
or FDA, in 1951. It deters
20:36
alcohol use by causing patients
20:39
to get ill when they drink. Naltrexone
20:42
was approved in 1984 and
20:45
Acamprosate 20 years later.
20:48
They help reduce alcohol cravings
20:51
and make withdrawal more manageable. These
20:54
drugs work fairly well, if not dramatically
20:56
well, in most patients, says
20:59
Joshua Lee of New York University
21:01
Grossman School of Medicine, who
21:03
specialises in addiction medicine. Other
21:06
drugs, such as Topiramate, are
21:09
used off-label for alcoholism,
21:11
that is, doctors prescribe them for reasons
21:14
beyond their approved use by the FDA.
21:18
These drugs are as effective for treating
21:20
alcoholism as selective
21:22
serotonin reuptake inhibitors,
21:25
such as Prozac, are for depression.
21:28
Yet they are rarely prescribed. Fewer
21:31
than 2% of patients with
21:33
alcohol use disorder report
21:36
using any medication. By
21:38
comparison, 22% of patients with
21:41
opioid problems and over half
21:44
of patients with depression take
21:46
prescription drugs. Why
21:48
the reluctance to use them? They
21:51
are not covered deeply in medical schools,
21:53
says Dr. Lee. Practitioners
21:55
lack adequate knowledge, training and
21:57
confidence.
21:59
Other common ailments, such as depression,
22:02
have more FDA-approved drugs,
22:05
enabling clinicians to mix and match
22:08
and deal with side effects. You
22:10
need as many medications as possible,
22:12
because some will work for some people,
22:15
but not others, says Lorenzo
22:17
Legio of the National Institute
22:19
on Drug Abuse and the National
22:21
Institute on Alcohol Abuse and
22:24
Alcoholism.
22:26
And many sufferers are loath to
22:28
seek help. Only 7%
22:30
get any treatment, be it counselling,
22:33
support groups or drugs. By
22:35
the time folks seek treatment, it
22:37
has been lingering for some time, says
22:40
Laura Ray, a clinical psychologist
22:42
at the University of California, Los Angeles.
22:46
Dr Ray estimates that many wait 10 years
22:49
to get help. Some may not
22:51
know that drugs are available.
22:54
All this could soon change. Some
22:57
aglutides, more commonly known as Ozempic
23:00
and Wigovi, have recently
23:02
become popular drugs to treat diabetes
23:05
and obesity. Ozempic
23:07
is not FDA-approved for weight
23:09
loss, but some doctors prescribe
23:11
it off-label. They may also
23:14
work for curbing drinking. If
23:16
approved for alcohol use disorder,
23:19
they could change how patients seek
23:21
help and how doctors respond.
23:25
Patients have reported reduced alcohol
23:27
cravings while on Ozempic
23:30
and Wigovi, and preclinical
23:32
trials on rats and monkeys suggest
23:35
that semaglutide helps reduce
23:38
such cravings. Of course
23:40
anecdotes and trials on animals
23:42
are not enough to declare them safe and effective.
23:46
The National Institutes of Health and
23:48
Oklahoma State University, or OSU,
23:51
have started randomized clinical
23:54
trials on humans to determine
23:56
if semaglutide is effective
23:59
for alcohol. This
24:02
drug could be a Prozac
24:04
moment for addiction medication, says
24:07
Cal Simmons, a pharmacology professor
24:09
at OSU, who is running the
24:12
clinical trial. Before
24:14
the FDA approved Prozac in 1987,
24:17
depression drugs were usually reserved for
24:19
the most serious cases due
24:21
to their side effects. With
24:24
Prozac, doctors became comfortable
24:26
prescribing drugs for depression. Doctors
24:30
started to ask for the medication. Similarly,
24:33
the popularity of Asempic and Wigovi
24:36
could change the way doctors and patients
24:38
think about drugs for alcohol addiction
24:40
if the FDA approves them
24:43
for that purpose. Dr
24:45
Simmons is optimistic about Simaglutide's
24:48
potential, but he cautions
24:50
patients and practitioners eager
24:52
to use it for alcohol use disorder
24:55
now. We don't know for
24:57
certain that it works, he says, and
25:00
his study's results will not be available
25:02
for about two years. Fortunately,
25:05
patients do not need to wait to get
25:08
help. They can turn to the drugs
25:10
for alcoholism that have been ignored
25:12
by many for so long.
25:19
I'm
25:28
Adam Roberts, and in London, this
25:31
is The Economist.
Podchaser is the ultimate destination for podcast data, search, and discovery. Learn More