The start of a decade is traditionally ushered in with a flurry of predictions. Those usually turn out to be wrong, but are illustrative of the expectations of the current age. Errors are informative, and show our limitations, our fads and misunderstandings. Looking back at predictions can be fun. Hindsight allows us to feel superior. We imagine that, if we could travel back in time, we could be of great use to our ancestors. Probably not. We would be chattering to them about scientific advances we did not fully understand, based on supportive technologies which would not be available, and without an accurate appreciation of what they already knew. They might find us arrogant, ignorant and pretty useless. For example, without looking it up, help your ancestors build a two way radio.
Prediction is not impossible: some bright and very diligent people can people can achieve accuracy by doing their homework on specific topics, as Philip Tetlock has shown. The trick seems to be to understand the background statistics in some depth, (say the number of times a government in power gets re-elected) and then make small adjustments in line with new circumstances, but rarely any dramatic ones. Business as usual seems to be a wise bet.
I have neither the wit nor the application to be a super-forecaster, but to reveal my current expectations, here are some wild surmises about the world in 2030.
1 Very few people will be using flying taxis, other than those already using conventional flying taxis.
2 Very few people will be being driven by autonomous vehicles.
3 Most cars will be powered by internal combustion engines.
4 Batteries will be only a little better than they are at the moment.
5 People all over the world will be richer and healthier.
6 Population growth will stabilize, except in Africa.
As you have seen, I can generate many testable predictions about very general subjects. Will I be any more capable if I make predictions about intelligence research?
I think the field will continue to flourish. “Ordinary” intelligence research, about what it is related to what in ordinary life, and how intelligence contributes to people’s achievements will continue to account for most published work. People will try to boost it by training techniques, and mostly fail. New ways will be found to test it, and turn out to contribute little. Some people will continue to doubt it can be tested, and is of any significance.
Brain scanning will improve, and with any luck will be carried out on larges samples of representative people, doing intellectual tasks which can be measured reliably, thus leading to better understanding about the links between brain activity and problem solving.
Artificial intelligence will be used as a tool to replace some routine pattern identifying tasks such as in medical screening, and in generating game playing strategies in simulations so as to design new materials and medicines. It will be the adjunct of choice when solving new problems. It will probably be used in almost every setting, and then become better understood and refined to deal with those tasks humans find too dull to dwell on for long.
Particularly, the study of genetics of intelligence will flourish. It is the best game in town. There is a risk of repetition, but further analyses will sharpen findings, and this will teach us more about our origins than has been known ever before. That is to say, by 2030 we will have learned more about our origins than previously known.
If debates about racial differences in intelligence are to be taken down to a genomic level, it will be necessary to assemble million person samples of non-Europeans, particularly Africans. It would be good if this were to happen in the next decade, but I assume such data will be collected in China, not Africa. Central governments have to be well-organised, positive about science and capable of delivering education and other basic public services for good statistics to be collected. Epidemiology tends to be a Scandinavian disorder.
I think there will be an increased acceptance of the screening of IVF embryos for health reasons, including for severe retardation from any cause. Journalists worry about “designer babies”. I think it far more likely we will move slowly towards “disease resistant humans”. As part of this process, if those few parents undergoing IVF choose the healthiest of their several possible embryos, there could be a boost of about 4 IQ points for those few children, which will be useful but hardly noticeable.
Currently, the moral issue revolves round embryo selection: given that a choice has to be made, what is the moral choice? “Do no harm” is the Hippocratic injunction. Is it moral to choose an embryo which will grow up to suffer a preventable disease? Most parents will probably feel that it is not, particularly if there is a family history of such a disease. Eventually, such families may feel that it is not moral to pass on the disease genes. Some children at maturity might start asking health related questions about the extent to which they were genetically screened, and be resentful if left carrying preventable genetic risks. This won’t happen this decade, but might become an issue for those born in the next one.
A justified concern about unintended effects, plus laws which prohibit such experiments, will prevent or at least delay attempts at genetic engineering. China has put He Jiankui and two colleagues into prison and banned them from receiving further research funding. It is unlikely that there will be any further CRISPR-cas9 type tinkering with human genomes for quite some while, and not by 2030. Work in other species will provide comparative data on the rate, extent and consequences of unintended effects. The rate estimates from mice is that they will be lower than 1% and usually minor in terms of consequences, but that is still high if contemplated for humans.
Looking further ahead beyond 2030, some people have made startling predictions about what genetic engineering might achieve in boosting human intelligence. I haven’t paid these much attention, arguing to myself that such speculations were premature and exaggerated. Using 4 IQ gains from embryo selection as the base rate, I saw little reason to get excited. Ever cautious, I thought I would check whether my caution was justified. I asked people in the business what might one day be possible, if it was both legal and possible to remove the currently known intelligence-reducing genes and ensure that everyone had all the currently known intelligence-boosting genes. To be clear, this estimate does not require further discovery of the links between the genetic code and human intelligence. It takes the current state of knowledge and assumes that gene editing techniques are good enough to go through the list of “intelligence genes” and make sure everyone has a full house.
Asking around, the answer seemed to be that CRISPR will probably not be an effective technique for making multiple changes in a single genome. So, the first point is that current technologies won’t work. CRISPR can handle short sequences (inserting and replacing) but not the longer ones required for the intelligence-boosting task.
Second, if one makes the leap of assuming that eventually there will be a new technique which can handle longer sequences, there is still a significant problem. Most of the hits mentioned in current research papers are not the causal sites, but merely correlated with the true causal sites. Much further work would need to be done to identify the causal sites. Only then can the manipulation of those sites be attempted.
Third, the potential overall boost to intelligence would need to be worked out in detail, causal site by causal site, and at the moment can only be estimated very roughly.
There would be an increase of about 10 standard deviations.
If that ever happens, all the other bets are off.
Enjoy the decade.